OpenAI’s new model broke rules and exploited loopholes more than any model METR has tested to date ...

AI is reducing obvious hallucinations, but deeper structural failures remain — and they may be far more dangerous for experts ...